Broker outage playbook: how to flatten without panic

[LondonScalper]

Platform freezes mid-London. Mobile app spins. VPS terminal is a white box. That’s not a “setup” problem — it’s an exit problem.

I keep a written outage playbook so I don’t invent one while stressed.

Pre-committed steps
1. Second path ready: webtrader / second device / phone app already logged in (test monthly).
2. Know the flatten path: one-click close all if it exists; otherwise hotkey sequence practiced when calm.
3. Broker phone / chat: number saved; what to say (“close all market positions on login X”) — no debate about the trade idea.
4. Risk assumption: if I can’t confirm flat in N minutes, I treat exposure as uncontrolled and stop adding for the day even after recovery.
5. Post-mortem: screenshot errors, timestamps, ticket IDs — for the broker and for my journal.

What I don’t do
• Open a “hedge” somewhere else without knowing net exposure.
• Refresh-spam until the session is a mess.
• Pretend it didn’t happen in the journal.

If you’ve had a real outage: what worked, what failed, and did the broker make you whole on obvious platform faults?

Process notes only — not advice.

[PropScalpDesk]

Outage: flatten path rehearsed monthly

Platform white-screen mid-session is an exit problem. I keep a written playbook so I do not design one while heart rate is elevated.

Pre-committed steps from this desk:

• Second path already logged in (webtrader or phone) — tested on a quiet Friday, not discovered during London
• One-click flatten / close-all known cold
• If both paths fail: broker desk number and account identifiers on a paper card, not inside the frozen VPS

I do not open new ideas while recovering connectivity. The only allowed action is risk to flat or confirmed reduce. Afterwards I log timestamp, symbol, and whether the mobile path worked — that log is what makes the next outage boring.

Prop accounts get the same playbook; arguing about a soft breach later is harder if you could not even flatten.

After any live outage I schedule a five-minute retro the same day: which path worked, which path lied, and whether the paper desk card was current. The playbook that is not rehearsed is fiction.

When did you last actually test the phone flatten path end-to-end, not just check that the app opens?

[Fairman]

Platform freezes mid-London. Mobile app spins.

Good list. The thing I'd put above all the flatten paths is making sure every open position already carries a stop that sits on the broker's server. A normal SL placed in MT5 or cTrader is held server side, so it still triggers if your terminal, VPS and phone all die at once. A stop managed by an EA or a hidden or virtual stop tool lives only on your machine and disappears with it. With real server stops, an outage becomes a known worst case rather than an open-ended one, and the playbook is about the remaining exposure, not panic. The setups where an outage really is uncontrolled are the ones trailing stops by hand or using hidden stops. Also worth keeping the account number and any security answers on paper, since the broker's phone desk will usually ask before closing anything.

[Shadow Trader]

One detail to add to Fairman's point on server stops. The trailing stop built into MT4 and MT5 runs in the terminal, not on the server, so if the VPS dies the trail freezes at its last level.

That is usually acceptable, since a frozen trail is still a stop. The real gap is an EA that was halfway through modifying a stop, and after one restart I found a position with no stop at all.

When I last timed the phone flatten from a cold start, it took 74 seconds. Most of that was a login code that went to an email inbox, and the mail client was open only on the machine that had frozen.

So the code now goes to the phone itself, and the paper card lists the account number for each login. During the monthly test I close one tiny demo position, rather than just checking the app loads.

A broker side outage is a different animal. If their server is down the stop waits too, and on reconnect it fills wherever price is, so the worst case is not quite as known as it looks.

Has anyone been made whole after a fill like that? I would like to know what evidence the broker asked for.

[Shadow Trader]

Pending orders are the part of an outage I underestimated. A buy stop sitting on the server keeps working when everything on my side is dead, so a frozen terminal can turn into a brand new position I cannot see.

That is why every pending order I place now carries its stop and target from the start. If it fills during an outage, it is at least a bracketed trade rather than open exposure.

Connection loss has been more common for me than a platform fault. My home line failed twice last year while the VPS ran fine, so the backup I needed was a phone hotspot rather than a second broker path.

Hedging accounts add a trap to LondonScalper's point about not opening a hedge elsewhere. On MT5 with hedging enabled, a sell placed from the phone to close a buy opens a second position unless you use close or close by.

In a panic that can look flat because the floating total stops moving, while you still hold two tickets and pay both spreads to unwind them. I practise the close from the phone app on demo for exactly this reason.

Has anyone found a broker that lets the client switch an account to close only? I have only seen the broker do it from their side.